Transcript Security of Electronic Transactions (Theory and Practice)

Security of Electronic Transactions
(Theory and Practice)
Jan Krhovják, Marek Kumpošt, Vašek Matyáš
Faculty of Informatics
Masaryk University, Brno
Outline
• Chip&PIN or signature cards – experiment
• Anatomy of the experiment
• Results of the Phase I & II
• Summary of both phases
• EMV standard
• Transaction processing
• Selected security mechanisms
• Other security implications
Security Protection of Information (SPI), May 2007
2/10
Anatomy of the Experiment
• The main question/objective
• Is it easier for an opportunistic thief to abuse Chip&PIN
or signature cards?
• We wanted to see/have experimental practical results
• First (warm-up) phase (over 40 people involved)
• Has taken place in pseudo-realistic conditions at a
university bookstore (Masaryk U., Brno)
• Age of the customers from 18 to 26 – “hired” students
• Time to practice a signature limited to about 30 minutes
• Time to practice “shoulder-surfing” about 2 hours
• Not genuine payment cards (no verification of the PIN)
• Second phase (over 35 people involved)
• Was realised in standard conditions
• In one of the largest supermarket in Brno
• Conditions set according to our experience from first phase
• Genuine payment cards (real verification of the PIN)
Security Protection of Information (SPI), May 2007
3/10
Results of the Phase I
• PINpad 1 (security/privacy shielding)
• Observers succeed in 6 from 17 PINs (35.3%)
• They needed one guess in 5 from the 6 PINs (83.3%)
• In 39 tips of 4-digit PINs (i.e., 156 digits)
• 75 digits guessed correctly (48%)
• PINpad 2 (no shielding)
• Observers succeed in 12 from 15 PINs (80%)
• Just one guess needed in 10 out of 12 PINs (83.3%)
• In 46 tips of 4-digit PINs (i.e., 184 digits)
• 129 digits guessed correctly (70.1%)
• Signatures (15+17 customers)
• Merchant detects 12 of 17 forging customers
• 5 forging customers passed (29.4%)
• 8 from 32 customers asked to sign twice
• After second signing: 4 detected & 4 passed
• We verified the signatures very carefully!!!
Security Protection of Information (SPI), May 2007
4/10
Results of the Phase II
• PINpads with & without security/privacy shielding
• 13 obs. from shielded pad, 7 from unshielded pad
• Observers succeed in 4 from 20 PINs (20%)
• 3 PINs from shielded pads, one from unshielded pad
• In 26 tips of 4-digit PINs (91 digits announced)
• 38 digits guessed correctly (42%)
• One (from three) observers group was more assertive and
able to closely follow the targets => best results
• Correct digits by groups: 25%, 27%, 68% (!)
• Third group: four correct PINs in 3 or less attempts
• Signatures (20 customers, stop after 17 succ. attempts)
• 10–30 minutes for practicing a given signature
• No problem reported from both the customers / till assistants
• No one was asked to sign again or to show an ID
• Some signatures were checked very poorly or not at all
Security Protection of Information (SPI), May 2007
5/10
Summary of Both Phases
• The signature forgers were newbies, as well as the
shoulder-surfers
• Correctly observed PIN digits (60% or 42%)
• Significant difference between fake signatures detection
(70% or 0%) – space for improvement
• Privacy shielding is really useful, however
• Majority of PINpads not equipped by shielding
• Light (=non-effective) privacy shielding in shops
• Some customers may have motoric difficulties
• Temporary remedy (?)
• Both PIN and signature
• Different PINs for low- and high-level transactions?
Security Protection of Information (SPI), May 2007
6/10
EMV (Europay, MasterCard, and Visa) Standard
• EMV 4.1 specification (4 books with ~800 pages)
• Interoperability & security of payment systems
• Smartcards, payment terminals, banking HSMs
• Introduction of Chip&PIN technology
• Magnetic-stripe cards can be easily copied (skimming)
• Transaction processing (online or offline)
• Authentication of on-card data
• Offline detection of fake (altered/duplicated) cards
• Static/dynamic data authentication
• Authentication of cardholders/users
• Based on handwritten signatures or PINs
• Priority list of card-supported verification methods
• Automatic risk analysis
• Online transaction authorization
Security Protection of Information (SPI), May 2007
7/10
Selected Security Mechanisms
• Static data authentication (SDA)
• On-card RSA signed static data
• Send to the terminal for offline verification
• No on-card asymmetric crypto => cheaper smartcards
• SDA still allows skimming
• Dynamic data authentication (DDA)
• Additional RSA key pair securely stored on card
• On-card signing of random challenge from terminal
=> more expensive smartcards
• DDA defeats skimming
• User authentication & card verification methods (CVM)
• CVM list included in signed data only optionally
• Adversary can modify this list of methods
• PIN => handwritten signatures
Security Protection of Information (SPI), May 2007
8/10
Other Security Implications
• Problem of the EMV specification
• Payment terminals protect interests of merchants
• Payment cards protect interests of banks
• Interests of customers are typically overlooked
• Malicious merchant can always cheat the customer
• Copying the cards
• Collecting of authorization data
• Relay the EMV protocol
• Electronic advocate
• Enters to the EMV protocol
• Protects only interests of the customer
• Portable electronic device between smartcard & terminal
• Can display the details of each transaction & accept/reject
Security Protection of Information (SPI), May 2007
9/10
Conclusions
• Introduction of Chip&PIN does not improve customer
protection against opportunistic thieves
• Factor in problematic repudiation of false transactions
• Good PINpad privacy/security shielding recommended
• The shop till clearly isn’t the right place to enter PINs
• Verification based on signatures
• Absolutely insufficient in a standard shop
• Better, e.g., in jewellery
• Shoulder-surfing is an underestimated issue
• Also in other (less “hostile”) environments, e.g., office
• Security of EMV systems still dependent on particular
implementation
• The system still protects mostly merchants and banks
Security Protection of Information (SPI), May 2007
10/10